[RELEASE] ScyllaDB 6.0.2

Release Notes
, ,
1

The ScyllaDB team announces ScyllaDB Open Source 6.0.2, a bugfix patch release of the ScyllaDB 6.0 stable branch. ScyllaDB Open Source 6.0.2, like all past and future 6.x.y releases, is backward compatible and supports rolling upgrades.

Users are encouraged to upgrade to 6.0.2.

Related links:

Issue fixed in this release:

  • Security: the new ScyllaDB image (GCP, AWS, Azure, Docker) includes a fix for OpenSSH Vulnerability CVE-2024-6387
  • Stability: Off-strategy compaction is used to make sstables conform to the compaction strategy after an operation such as repair. Off-strategy compaction for TWCS will now have less storage space overhead. #16514
  • ADMIN API: using the admin REST API may lead to a crash with assertion `!_end && !_zc_bufs && “Was this stream properly closed?”’ failed #19494
  • Stability: Batches (as generated by the BATCH statement) are held in the system.batchlog table. As this table can accumulate a lot of tombstones, we now take steps to ensure these tombstones are purged eagerly. This is important for repair, which replays the batch log. #19376
  • Packaging: Bundle cqlsh with support for “DESC SCHEMA WITH INTERNALS” #18989. Also see #19334 below.
  • CQL: DESC does not report CQL Table Extensions#19334
  • Stability: chunked_vector::reserve_partial does not work as documented if size > max_chunk_capacity() #19254
  • CQL: Disabling tablets using ALTER KEYSPACE should be forbidden #18795. This was already non supported, but now blocked in the CQL level.
  • Performance: Statements such as SELECT count(*) use an internal map-reduce service to parallelize the query. ScyllaDB no longer does so for single-partition queries as they don’t benefit from it. #19349
  • Stability: Failed to add node in parallel #19523
  • Stability: hinted handoff: Scylla may initialize endpoint managers for invalid hint directories #19172
  • Stability: init - Startup failed: std::runtime_error (Failed to mark node as alive in 30000 ms …). The root cause is a node booting in gossip topology waits until all NORMAL nodes are UP. If we removed a different node just before, the booting node could still see it as NORMAL and wait for it to be UP, which would time out and fail the bootstrap. #17526
  • Topology Update: Scylla allows replacing a node in one dc/rack with a node in a different dc/rack which requires full cluster repair and then cleanup all nodes. This is now blocked, and you should start a new node and decommission the old one instead. Use Replace operation in the same DC/Rack only. #16858
  • Storage: ScyllaDB now use the latest ME SSTable format by default #18995
  • Stability: tablet storage is not freed when migrating a tablet replica away #16946
  • Stability: tasks: abort method hits use after free #19304
  • Performance: The gossiped view update backlog is only read on shard 0 #19232
  • Stability: View update generation times out during index building, wich result is redundant fake warnings. #19261
1 Like